What are you working on?
Michael Koppelman
if so, what they are working on. Anyone care to share?
M.
James Roe
I'm curious who on the list is working on data mining right now and,
if so, what they are working on. Anyone care to share?
I've had trouble with the data mining concept, especially as it has been described in the past two workshops. I finally figured it out (I think) and you put your finger on it here. I can see and understand that IF you have identified a star (or whatever) of interest how data mining (as described in the workshops) is used. But my desire is to FIND an interesting project in the sea of available information. My crude metaphor is a walk in the woods. I might stumble across a patch of interesting wild flowers that should be written up for the record. It might be a hole in the ground that might house an interesting creature if it could just be observed enough. Does this make sense?
On the other hand, I did have an opportunity recently to do the more familiar form of data mining. I provide telescope time and data gathering for students at our local community college who are taking an astronomy lab course. One of them had the idea to get a complete light curve for a short period DSCT star V0457 Oph. I checked and there was no data in the AAVSO database. We spent a couple of sessions and ended up getting a complete light curve with both halves separated by 7 periods. I uploaded the data to the DB and then went out and searched the ASAS3 data base. Sure enough, there were several years of data there but the cadence being every few days or so no periodicity could be perceived in the LC. Our data showed a very respectable phase diagram, but the addition of the ASAS3 data really nailed it. After some discussion about the value of adding ASAS3 data o the DB (on the photometry discussion group), I went ahead an uploaded the ASAS3 data. It was fun.
Jim Roe [ROE]
Doug Welch
Folks,
I thought that I would mention a specific instance of data-mining that
worked out well for me and my collaborator, Geoff Clayton. We were
searching for R CrB candidate lightcurves in the MACHO databases
containing LMC, SMC, and Bulge data. Needless to say, there was no
lightcurve flag that said "I'm a previously unrecognized R CrB star". So
our trick was to define statistics and narrow down the range of
combinations of such statistics to produce a list of candidate stars
which could then be examined in a finite amount of time by eye. Skewness
was one such statistic - many (but not all!) R CrB spend a lot of time
at maximum light, so non-maximum brightness points tend to fall on one
side of most of the points. That is also true of eclipsing binaries, of
course. So one needs to try combinations of statistics. They also need
to be quick to evaluate (or be worth the trouble of defining for all
stars once and for all!) Most eclipsing binaries would be shorter-period
and so would have many minima during our 8 years of data-taking. The
very long-period eclipsing binaries are interesting in their own right,
so we didn't mind finding those as R CrB "contaminants"!
This all resulted in papers like:
http://adsabs.harvard.edu/abs/2005AJ....130.2293Z
http://adsabs.harvard.edu/abs/2001ApJ...554..298A
http://adsabs.harvard.edu/abs/1996ApJ...470..583A
I would say that one of the current impediments to making better use of
existing stellar photometry databases is the lack of supplemental,
easily-defined, summary (or "aggregate") statistics such as skewness and
kurtosis. If the database is "live", these would need to be recalculated
every so often (although a yearly set of such indices could be easily
searched for changes, too!)
For folks who work with Postgresql, I have defined aggregate functions
(just modifying existing code for the standard deviation calculation!)
that I would be happy to provide to interested parties.
Finally, I would remark that one of the things I noticed with the MACHO
data was how easy it was to classify many stars simply by eye-balling
the lightcurve over the entire duration - that is, completely before
any time-series analysis. The way the points fill the space is quite
distinctive - you can tell a sawtooth curve from a sinsusoid, from a
"bump" pulsator - even from a multimode star. All this without any
time-consuming, time-series analysis. This, of course, is just a
restatement of my previous paragraphs saying that skewness and kurtosis
tells you a lot, especially combined with amplitude, color, ... and
numbers that likely *do* exist in the summary statistics already.
Enough for now!
Doug
Martin Nicholson (NMR)
this article
http://var.astro.cz/oejv/issues/oejv0102.pdf
However ... while the "usual culprit" can access the group and so
feels free to post (elsewhere) things like "Can people even imagine
someone setting up a group to help people do data mining and not
having a simple project ready for teaching any interested newbies?
Usual thing, AAVSO needs volunteers to do it for them, not sure where
the credit will go, but you know, who cares." well then I am not very
motivated. Sorry but I did warn folk about this possibility some time
ago - in VSNET-CHAT to be precise.
Set the group up so that only group subscribers can read or make posts
and I would like to get involved but until then ....!
Martin Nicholson
Daventry, UK (AAVSO - NMR)
Michael Koppelman
Either contribute to the group or do not, its your choice. This group
will remain open and I consider that a good thing.
Cheers,
M.
On Jul 29, 5:34 am, "Martin Nicholson (NMR)"
Patrick Wils
The start of the ASAS light curve (at minimum) coincides with the last point in the infrared data of Whitelock et al. (2006): <http://cdsarc.u-strasbg.fr/viz-bin/vizExec/Vgraph?cat=J%2FMNRAS%2F369%2F751%2F.%2F09164-5349&-x0=49900&-x1=52100>. 2MASS probably caught it during the decline to the last minimum (J=5.75 at JD 2451550).
The low amplitude at longer wavelengths is also typical for RCB stars.
Patrick
Doug Welch
Nice find, Patrick! That one would be one of the dozen brightest known -
at least at maximum light!
By what path did you uncover it?
Cheers,
Doug
Patrick Wils
Nothing in particular really, just looking at lots and lots of light curves.
Patrick
aaron price
exploratory data analysis of the evaluation component of the citizen
sky user registration process. You should see some of the answers
we've gotten to the science philosophy questions! :)
I'm polishing a simple enough paper on BZ UMa that I hope to submit
to PASP shortly after the workshop.
My next astro project hopefully will be to use the AID to characterize both the
precision and accuracy of visual observations. Basically, a follow up
on this poster I did a few years ago:
http://www.aavso.org/vstar/aas-jan07-visualprecision.pdf
I have ideas on how I'd like to do it, but right now I don't plan to
begin until I finish my dissertation (hopefully, December 2010). But
if anyone wants to participate and get started sooner, send me an
e-mail or talk to me at one of the upcoming meetings.
I also have long had another idea: Amateurs have long reported flares of mira
stars on the order of a few seconds to a few minutes. But no one has
been able to prove it (Dale Mais tried gallantly). My suspicion is
that such flares have been
flagged in the AID as "discrepant". So we could do a statistical
analysis of the rate of discrepant observations vs. spectral type.
It's a little more complicated than it sounds since you need to
control for changes in magnitude, chart sequences and other sources of
discrepant data. Hence why this would likely have to come some time
after the first project is completed, as that project will provide the
data we need to establish some of the covariants in the stats. Again, happy
to work with anyone on this.
Aaron
Aaron
--
Aaron
Simonsen
Greaves should be banned from posting or joining this or any other
AAVSO discussion list, based on past and current behavior. There is no
reason left to sympathize with or accept either of them as having any
redeeming value. The end result will always end up on the negative
side of the ledger.
Nicholson has shown the poor judgment to post a sideways attack on
Greaves as his first "active contribution" to this list, knowing full
well that this would precipitate yet another disgusting round of non-
reply nasty spam emails from Greaves, and lengthy flame wars on other
public discussion groups who haven't had the intelligence or taste to
ban these two yet.
Nicholson has no qualms about instigating or encouraging further long-
winded Greaves attacks on the AAVSO and individuals by remarking and
answering Greaves attacks publicly. He's even gone so far as to create
web pages filled with negative remarks, selected quotes from other
discussions and snide personal attacks dressed up as defenses against
his being wronged.
That only makes the situation worse and worse, the end result being
several thousand words of negative attacks on our organization, the
projects we are trying to develop and the people who are trying to do
something positive in the face of this wilting negativity.
Its like the village idiot, poking and poking and poking at the
hornets nest, and then acting like he's surprised when he gets stung.
Fine, be a dope. But I'm tired of being stung because of Nicholson's
ignorance and Greaves' blind rage.
It is a sad day when a Google search on 'AAVSO data mining section'
brings up Greaves' and Nicholson's flame wars higher in the results
than the actual section web pages or discussions. In turn, these
search results link out to yet more attacks from them on each other
and the AAVSO. The net result is that this section, in its nascent
form, is drowned under a tidal wave of negativity and hatred created
by these two.
I have banned them both from every other list, discussion, blog or
group I administer because they cannot be trusted to behave civilly.
I don't control this list, so I can only vote with my participation. I
will not subscribe to this list if these two are not banned from
posting.
Furthermore, I think it is time for council to examine whether we want
anyone who builds hate sites and recirculates negative flame war
discussions to the internet signing his name as (AAVSO-NMR).
Mike Simonsen
(AAVSO-SXN)
On Jul 30, 1:27 pm, Michael Koppelman <michael.koppel...@gmail.com>
wrote:
Pedro Pastor
I am working on Pattern Recognition (as a researcher at my University). I've been developing projects that has nothing to do with astronomy (some of them very theoretical), but I am recently involved in a very practical project dealing with "Automatic Star Spectra Classification".
On the other hand, I'm an active photometrist collaborating with some professional researchers. Some times I think of applying the Pattern Recognition techniques I know to the field of Variable Stars Light Curve Classification, BUT I've usually found difficult to take control on the great variety of catalogues (I mean, formats, accesibility, coverage, quality of the data, and so on), and that point has prevented me from enter into this interesting problem.
Maybe, with the help from people on this grup I could learn a bit more on the use of such catalogues and I dare to enter into the job.
Note: Yes, I know, in the case of star Spectra, the problem on using catalogues in even more difficult to deal with, but I'm fortunate. I'm working using a "controled set" of spectra produced by a specific research team at the IAC ("Instituto de Astroficsica de Canarias").
Best regards,
Pedro
Wolfgang Renz
Well sad.
While I know that JG could make usefull new contributions, I doubt
this with MN. Anyway, in the general mood both are, I don't think that
any group could benefit from them without risking a never ending
flame war. I also don't want to see MN regularly advertising any of
his private pages here as long as he creates and links his hate
blogs and hate pages.
Clear skies
Wolfgang
--
Wolfgang Renz (AAVSO.RWG), Karlsruhe, Germany
Doc Kinne
If I may be so bold as to bring this thread back on track...
I am in the midst of taking a look at three stars - LX Cyg, W Dra, and
BH Cru - for fundamental period changes over time.
This is not groundbreaking research by any means. Matt Templeton has
exactly this sort of paper in the pipe for T UMi and he, Lee Ann
Willson, & Grant Foster published such a paper on RU Vul last year.
However, using the data for these stars in the AID, and such tools as
TS, WWZ and CLEANest, I am hoping to give myself some very much needed
experience in statistical data analysis, and make a tiny contribution
that will lead to something published.
About a year ago I attempted to help Aaron with the data analysis of his
study of BZ UMa superhumps. I did a good bit of work, but never found
the superhumps. Obviously Aaron did and so I'm also going through his
paper, and my analysis documentation, to see "where I went wrong."
I may post from time to time my progress as an example of a blind man
stumbling in the dark. So, if you see me walking around all bloody from
running into things, bandages would be appreciated. :-)
--
Richard "Doc" KINNE, BA, MSc., AMAAS
<kinnerc @ gmail . com> (sent from Linus)
mark_h37
confirmation of possible companion stars to the stars listed in Tom
Cragg's 1963 paper regarding. Flat-Bottomed Light Curves of LPV. So
far I am about half way through the list and have not found any as of
yet. I have also checked a few of them in the Washington Double Star
Catalog with out any success as of yet. It would be nice to get at
least one positive confirmation so that I felt like I was on the right
track, but such is life.
On Jul 29, 1:19 am, Michael Koppelman <michael.koppel...@gmail.com>
wrote:
Arne Henden
with good results. The trick is to look at the imagery, and not at
Simbad. What you want is to find high resolution images taken when
the LPV was near minimum. You want something with arcsec-like
resolution.
That leaves out NEAT, CRTS, and to some extent, the photographic
plate surveys like POSS. Your best bets are the image archives
kept by the various observatories, such as the Isaac Newton Group
(which has the WFC survey as well as IPHAS), CFHT, etc. - you want
wide field with lots of pixels. An alternative is to post the list
of objects, and let observers in good sites take some images to look
for companions. I can do that to some degree with the AAVSOnet
scopes,
or with observing proposals at better sites. So - you can either
datamine the existing archives to find the right images, or datamine
to find likely objects and provide input to others.
I think this is a good project and would like to see someone make
progress!
Arne
mark_h37
Thank you for the information. Below is a copy of what I had sent Dr.
Matthew Templeton. As can be seen you are right there is little
support for this in the sinbad data base. I started there because it
was what I was familiar with. Since this is the beginning of the
semester for me my spare time is at a minimum but as things settle
down I will be attempting to look into some of the other data bases as
well. I have put together an updated list of these stars that I am
going to pass on to David Boyd to see if he can image some of them
using the 2m Faulkes Telescopes that he has access to.
Mark
I did a search of the Simbad Astronomical Data base and found two
additional stars on the list that are double stars. They are:
Y Vel CCDM J09290-5211A
ST Sgr CCDM J19016-1247AB, IDS 18559-1255
These where the only two on the list from Tom Cragg's paper that were
listed as double stars. So far this is all that I have found related
to this.